The classification of Brain tumor, Lung, and Colon Cancer plays a crucial role in the diagnosis of diseases such as brain tumors, lung cancer, and colon cancer. These images require precise interpretation due to their complex and microscopic patterns, which are often difficult to analyze manually. To solve this problem, a hybrid deep learning model combining EfficientNetB7 and bidirectional LSTM (BiLSTM) layers is proposed. EfficientNetB7 is used for feature extraction as it can capture complex spatial patterns, while BiLSTM layers are used to analyze the feature vectors' temporal dependencies. Transfer learning with pre-trained EfficientNetB7 weights speeds up training and improves accuracy, while data augmentation enhances the model's generalization ability. The proposed model achieves a test accuracy of 98.17%, a mean precision of 98.23%, a mean recall of 98.50%, and a mean F1-score of 98.36%, demonstrating its reliability and robustness. An analysis of the confusion matrix also shows minimal misclassification, confirming the model's potential for real-world applications in medical diagnostics.